The Reserve Bank of Australia The process

Monetary policy in Australia is set by the Reserve Bank of Australia (RBA) Board. This Board consists of nine members, seven of whom are external to the Bank. No regular set of published forecasts is provided but they can sometimes be gleaned, at least on a qualitative dimension, from studying the Semi-Annual Statement on Monetary Policy, the quarterly reports on The Economy and Financial Markets and speeches of RBA officials.12 One peculiarity of monetary policy in Australia is that the Secretary of the Treasury sits on the Board. Perhaps this is responsible for another unusual arrangement in the form of a Joint Forecasting Group, which involves both the Treasury and the RBA forecasters along with representatives of other Commonwealth Government Departments with input into economic policy formation. Although there is no requirement that a single forecast should be adopted by all members of this group, it would be surprising if the differences in forecasts are very large. The presence of the Joint Forecasting Group means that concentrating upon the forecasting activities at the RBA in isolation may be a little misleading regarding the overall forecasting system.

The RBA constructs forecasts using several of the approaches described earlier and these are presented to internal staff meetings in the lead-up to each Board meeting. A small core macroeconometric model is used to produce a set of no-policy change forecasts, adjusted in light of special anticipated events using intercept corrections. For example, inflation forecasts for the year 2000 have to recognize the impact of the pending introduction of a Goods and Services Tax. The model-based forecasts are then compared to those from the groups responsible for producing separate forecasts of real and nominal variables. These groups use a variety of spreadsheet or indicator models, as well as incorporating their own judgment and survey data. For example, survey data on expected future capital expenditure by the private sector in Australia have proved to be quite accurate predictors of investment. There is often conflict between the projections of the "spreadsheet" and "core modeling" groups and this is resolved by the former deciding whether to adjust their forecasts in light of the those produced by the latter. The forecasts of both the modeling and "spreadsheet" groups are then passed up to senior management who impose another round of judgment in reaching a final forecast. Sometimes these adjustments are quite large owing to a feeling that there is too much inertia in the forecasts that come out of the indicator framework. After all these modifications, a final set of forecasts is sent to the Board for use in its deliberations.

The process just described is similar for both short- and long-term forecasts. Unlike some other institutions, the model doesn't become more influential as the forecast horizon lengthens, although long-term forecasts do tend to revert toward trend values that are found in historical data. Such long-run convergence is very common among forecasters. A striking example was recently afforded when the 1999 benchmark revisions to U.S. real GDP data raised the average historical growth rate by 50 basis points. Almost immediately, most forecasters raised their long-term forecasts of U.S. growth by exactly this amount.

Overall, although both indicator models and a core model are used in the RBA for forecasting purposes, there is extensive modification of their output through the judgment of higher-level bank officials, albeit in a way that is hard to systematize. Certainly, there is nothing like either the Bank of Canada or Reserve Bank of New Zealand approach to the task. Simulations are sometimes run with the core model to compute the optimal policy path consistent with that model and this is taken as a useful check on policy actions that are either proposed or have been taken. With that exception, it seems rare for the RBA to formally study different policy scenarios, and that may be an important part of the explanation for why the core model has a much smaller role in the policy process compared to the situation in other central banks.

0 0

Post a comment